Vice President of Emergent AI

Max Versace | Analog Devices

KEYNOTE: "Edge AI is not a place. It’s a constraint". Edge AI is often framed as “running models on devices”, but this often distracts from the real issue, which is not where AI is running, rather, what real-world constraints it faces. As an example, closing real-world sensory-motor loop in a humanoid robotic hand faces latency, power, bandwidth, AI workloads orchestrations and trust issues, and requires AI hardware that are humble enough to satisfy these constraints, as well as AI models that can face the tough reality of our physical world without breaking. This keynote reframes edge AI as a system discipline, using the brain as the reference design for Physically Intelligent sparse, event-driven, predictive computation. We connect those principles to modern edge architectures, and dive into and dexterous robotics example where tactile sensing and control demand sub‑millisecond, low power local intelligence. In Physical Intelligence at the Edge, Sensing is AI, and the next edge stack will fuse sensing–compute–learning into deployable physical system that are designed to solve real-world constraints.